Probabilistic assessment of vegetation vulnerability to drought stress in Central Asia.

The increasing frequency and intensity of droughts in a warming climate are likely to exacerbate adverse impacts on ecosystems, especially for water-limited regions such as Central Asia. A quantitative understanding of the impacts of drought on vegetation is required for drought preparedness and mitigation. Using the Global Inventory Modeling and Mapping Studies NDVI3g data and Standardized Precipitation Evapotranspiration Index (SPEI) from 1982 to 2015, we evaluate the vegetation vulnerability to drought in Central Asia based on a copula-based probabilistic framework and identify the critical regions and periods. Furthermore, a boosted regression trees (BRT) model was also used to explore the relative importance of environmental factors and plant traits on vegetation response to drought. Additionally, we also investigated to what extent irrigation could alleviate the impacts of drought. Results revealed that months from June to September was the critical period when vegetated areas were most vulnerable to drought stress. The probabilities of vegetation loss below 20th quantile under extremely dry in these months were 68.7%, 69.4%, 71.0%, and 67.0%, respectively. Regarding vegetation-vulnerable regions, they shifted with different growth stages. During the middle of the growing season, semi-arid areas were the most vulnerable regions, whereas the highest drought-vulnerable regions were observed in arid areas during other periods. The BRT results showed that plant traits accounted for a large fraction (58.9%) of vegetation response to drought, which was more important than ambient soil environment (20.8%). The analysis also showed that mitigations from irrigation during July to September were smaller than in other months. The results of this paper provide insight into the influences of drought on vegetation and may contribute to drought mitigation and land degradation measures in Central Asia under accelerating global warming.

Assessing vegetation stability to climate variability in Central Asia.

The dramatic climate change has far-reaching impacts on vegetation in drylands such as Central Asia. Recent attempts to assess vegetation stability to short-term climate variability often account solely for vegetation sensitivity or resilience but ignore the composite effects of these two indicators. Meanwhile, our understanding of the vegetation stability at the seasonal scale remains insufficient. In this study, considering the cumulative effects of vegetation response to three key climate factors, we assessed the stability of vegetation in Central Asia using normalized difference vegetation index (NDVI) and the meteorological data from 1982 to 2014 by integrating vegetation sensitivity and resilience, and further identified the critical regions and seasons of vegetation that experience high risks of pending change. The results show that the sensitivity of vegetation has a strong correlation (R2 = 0.83, p < 0.001) with the aridity index (AI), with the vegetation of drier areas having lower sensitivities to climate variability. At the temporal scale, the sensitivity of vegetation to climate variability varied among different seasons. The average vegetation sensitivity index (VSI) is 41.17, 33.32 and 28.63 in spring, summer and autumn, respectively. Spatially, a trade-off between vegetation sensitivity and resilience is found both for the growing season (R2 = 0.67) and seasonal scale (R2 = 0.71, 0.32 and 0.43 for spring, summer and autumn, respectively), regions with high vegetation sensitivity were always accompanied by strong resilience. Based on the relationship between vegetation sensitivity and resilience, we further identify the critical regions and periods of vegetation with high change risk in Central Asia. Results suggest that herbaceous plants in semi-arid areas present high instability, especially in summer. This study offers a comprehensive perspective to assess vegetation stability to climate variability and the results will facilitate the protection of ecosystems and the implementation of sustainable development goals in Central Asia.

Exploring Variability in Landscape Ecological Risk and Quantifying Its Driving Factors in the Amu Darya Delta.

Examining the drivers of landscape ecological risk can provide scientific information for planning and landscape optimization. The landscapes of the Amu Darya Delta (ADD) have recently undergone great changes, leading to increases in landscape ecological risks. However, the relationships between landscape ecological risk and its driving factors are poorly understood. In this study, the ADD was selected to construct landscape ecological risk index (ERI) values for 2000 and 2015. Based on a geographically weighted regression (GWR) model, the relationship between each of the normalized difference vegetation index (NDVI), land surface temperature (LST), digital elevation model (DEM), crop yield, population density (POP), and road density and the spatiotemporal variation in ERI were explored. The results showed that the ERI decreased from the periphery of the ADD to the centre and that high-risk areas were distributed in the ADD's downstream region, with the total area of high-risk areas increasing by 86.55% from 2000 to 2015. The ERI was spatially correlated with Moran's I in 2000 and 2015, with correlation of 0.67 and 0.72, respectively. The GWR model indicated that in most ADD areas, the NDVI had a negative impact on the ERI, whereas LST and DEM had positive impacts on the ERI. Crop yield, road density and POP were positively correlated with the ERI in the central region of the ADD, at road nodes and in densely populated urban areas, respectively. Based on the findings of this study, we suggest that the ecological constraints of the aforementioned factors should be considered in the process of delta development and protection.

Disentangling the relative impacts of climate change and human activities on arid and semiarid grasslands in Central Asia during 1982-2015.

In recent decades, climate change and human activities have severely affected grasslands in Central Asia. Grassland regulation and sustainability in this region require an accurate assessment of the effects of these two factors on grasslands. Based on the abrupt change analysis, linear regression analysis and net primary productivity (NPP), the spatiotemporal patterns of grassland ecosystems in Central Asia during 1982-2015 were studied. Further, the potential NPP (NPPP) was estimated using the Thornthwaite Memorial model and the human-induced NPP (NPPH), which was the difference between NPPP and actual NPP, were used to differentiate the effects of climate change and human activities on the grassland ecosystems, respectively. The grassland NPP showed a slight upward trend during 1982-2015, while two obvious decreasing periods were found before and after the mutation year 1999. Additionally, the main driving forces of the grassland NPP variation for the two periods were different. During 1982-1999, climate change was the main factor controlling grassland NPP increase or decrease, and 84.7% of grasslands experienced NPP reduction, while the regions experiencing an increase represented only 15.3% of the total area. During 1999-2015, the areas of increasing and decreasing grassland NPP represented 41.6% and 58.4% of the total area, respectively. After 1999, human activities became the main driving force of the NPP reduction, whereas climate change facilitated grassland restoration. The five Central Asian countries showed widely divergent relative impacts of climate change and human activities on NPP changes. In Uzbekistan and Turkmenistan, anthropogenic decreases in grassland NPP intensified during 1982-2015, while the negative anthropogenic effects on grassland NPP in Kyrgyzstan and Tajikistan moderated. Further analysis identified precipitation as the major climatic factor affecting grassland variation in most areas of Central Asia and overgrazing as the main form of human activity accelerating grassland degradation. This study improves the understanding of the relative impacts of climate change and human activities on grasslands in Central Asia.

Monitoring land sensitivity to desertification in Central Asia: Convergence or divergence?

In Central Asia, desertification risk is one of the main environmental and socioeconomic issues; thus, monitoring land sensitivity to desertification is an extremely urgent issue. In this study, the combination of convergence patterns and desertification risk is advanced from a technical perspective. Furthermore, the environmentally sensitive area index (ESAI) method was first utilized to monitor the risk of desertification in Central Asia. In the study, the spatial and temporal patterns of desertification risk were illustrated from 1992 to 2015 using fourteen indicators, including vegetation, climate, soil and land management quality. The ESAI spatial convergence across administrative subdivisions was explored for three time intervals: 1992-2000, 2000-2008 and 2008-2015. The results indicated that nearly 13.66% of the study area fell into the critical risk of desertification from 1992 to 2008. However, the risk of desertification has improved since 2008, with critical classifications decreasing by 19.70% in 2015. According to the mutation year detection in the ESAI, 25.89% of the pixels with mutation years from 1992 to 2000 were identified, and this value was higher than that during the other time periods. The convergence analysis revealed that the desertification risk for 1992-2000 tended to diverge with a positive convergence coefficient of 0.13 and converge over the 2000-2008 and 2008-2015 time periods with negative convergence coefficients of -0.534 and -0.268, respectively. According to the spatial convergence analysis, we found that the divergence patterns in northern Central Asia from 1992 to 2000 resulted from the effects of the Soviet Union collapse: cropland abandonment in northern Kazakhstan and rangeland abandonment in Tajikistan, Kyrgyzstan and eastern Kazakhstan. In contrast, most areas from 2000 to 2008 experienced increased sensitivity to desertification with the convergence pattern caused by decreased precipitation, especially in northern Central Asia. However, convergence patterns were found in most regions for 2008-2015 with regard to augmented precipitation, which resulted in decreased sensitivity to desertification. Moreover, the low sensitivity areas were more likely to converge under increased precipitation. In this region, the findings of our study suggested that spatial convergence and divergence acted as related predictors of climate change and human activities, respectively. Thus, the ESAI convergence analysis was considered to provide an early warning of potential desertification.